While this raises a host of thorny issues, it also would accomplish one of the critical steps in getting people to buy into assessment – the data would be usable and, more importantly, _used_.

(Although it does remind me of the professor in library school who claimed that no student who got a B or higher in her course had ever failed their comps. I got my A, and I passed my comps, but it’s a pretty weird correlation to hang your hat on…)

If I were to teach an Econ course now I’d evaluate the students in two quite different ways, one as understanding models (and then depending on how advanced the course also on building models of their own) and two on telling stories that relate the models to known economic phenomena. Engineering students have a reputation about being good with the models but bad with the stories. Many Business students are just the opposite. You really need multiple ways of thinking about the economics and I suspect also about many other disciplines.

One other point. Accreditation is a reality as are grades (in my ideal world both would go away, but I still believe in the tooth fairy) so there is the real issue of whatever assessment we come up with can it be aggregated in some meaningful way and can we make comparisons across students and across cohorts of students? For both of these some Likert-style indicators are useful, a necessary evil if you will. I liked the participant perception indicator that Carl Berger show use some years back.
http://sitemaker.umich.edu/ppi.evaluation.tool/home
It is do-able, a definite plus, can be used in a longitudinal way, and does recognize that multiple dimensions in assessment can be better than reducing everything down to a single dimension. It does measure attitudes, not performance. The latter, as you say, is messy to assess in the meaningful cases. For that I’d prefer to give written feedback. Maybe we need both.

Although it is very difficult to assess “education”, it is quite easy to assess if you are meeting the goals of the curriculum. Yes, very few teachers even make an effort to do so.

Assessment in a Web2.0 environment is not that different. Web2.0 provides a platform for different assessment methods, activities and reports. But assessment remains quite unchanged.

We should never really worry too much about complexity. Think of how complex a task it would be to properly compensate someone for their time given that we live only for a short while. Yet something as simple as money does the trick.

Gilbert

The underlying problem, though, as you suggested in your talk, is that learning is not about valuation. It’s not about the knowledge a student can recite. Give me a reliable theory of cognition and maybe we can get to a theory of learning, and maybe, just maybe, we can figure out a way to measure learning. All we do now is some fancy foot work of what I, as a teacher, can represent, through grades, about what I know about what students know at a certain point and time.

There’s no way for anyone to assess what I have learned from listening to your talk. I can’t do it myself. So all we are left with is the ethical obligation to learn from our experiences and seek improvement (i.e. to assess).

In order to create accurate assessment, it would seem to be essential to know where education should go in the coming decade. Everyone who can predict the future please hold up their hands. Anyone?